An approach to dynamic hand gesture modeling and real-time extraction

In the process of vision-based hand gesture human­computer interaction, to extract the motion hand gesture accurately and quickly is the premise and key of improving recognition hand gesture's speed and accuracy. This article presents a method of a new dynamic hand gesture modeling and real-time extraction for hand gesture feature, combining the three factors — starting hand posture, intermediate trajectory and ending hand posture. Firstly this method determines the starting and ending of dynamic hand gesture by detecting the change in movement rate. Then it breaks up starting hand posture and ending hand posture based on motion and color information and real-time tracks the trajectory for hand gesture by using Mean shift algorithm combining target detection. The simulation shows that the approach has good real-time and robustness in the extraction for motion hand gesture.